University of Nebraska - LincolnDigitalCommons@University of Nebraska - Lincoln

Publications from USDA-ARS / UNL Faculty USDA Agricultural Research Service --Lincoln,Nebraska

1-1-2013

Insect communities in soybeans of eastern SouthDakota: The effects of vegetation management andpesticides on soybean aphids, bean leaf beetles, andtheir natural enemiesJonathan G. LundgrenUSDA-ARS, jonathan.lundgren@ars.usda.gov

Louis S. HeslerUSDA-ARS

Sharon A. ClaySouth Dakota State University

Scott F. FaustiSouth Dakota State University

Follow this and additional works at: http://digitalcommons.unl.edu/usdaarsfacpub

This Article is brought to you for free and open access by the USDA Agricultural Research Service --Lincoln, Nebraska at DigitalCommons@Universityof Nebraska - Lincoln. It has been accepted for inclusion in Publications from USDA-ARS / UNL Faculty by an authorized administrator ofDigitalCommons@University of Nebraska - Lincoln.

Lundgren, Jonathan G.; Hesler, Louis S.; Clay, Sharon A.; and Fausti, Scott F., "Insect communities in soybeans of eastern SouthDakota: The effects of vegetation management and pesticides on soybean aphids, bean leaf beetles, and their natural enemies" (2013).Publications from USDA-ARS / UNL Faculty. Paper 1164.http://digitalcommons.unl.edu/usdaarsfacpub/1164

Insect communities in soybeans of eastern South Dakota: The effects of vegetationmanagement and pesticides on soybean aphids, bean leaf beetles, and theirnatural enemies

Jonathan G. Lundgren a,*, Louis S. Hesler a, Sharon A. Clay b, Scott F. Fausti c

aUSDA-ARS, North Central Agricultural Research Laboratory (NCARL), 2923 Medary Avenue, Brookings, SD 57006, USAb Plant Science Department, South Dakota State University, Brookings, SD 57007, USAcDeparment of Economics, South Dakota State University, Brookings, SD 57007, USA

a r t i c l e i n f o

Article history:Received 28 February 2012Received in revised form8 August 2012Accepted 9 August 2012

Keywords:Biological controlCover cropIntegrated pest managementOrius insidiosusPredatorSystemsWeed

a b s t r a c t

Although most pests of soybeans, Gycine max (L.), in the Northern Great Plains are managed usingpesticides, farm management practices that encourage biodiversity offer promising long-term, sustain-able solutions for controlling insect and weed pests profitably. The recent invasion of the Northern GreatPlains by the soybean aphid (Aphis glycines Matsumura; Hemiptera: Aphididae) has had potentiallyimportant implications for insect communities in soybeans of this region, although recent descriptions ofthis regional community are scarce. We describe how three pest management systems that vary in theintensity with which they rely on herbicides and insecticides (chemically intensive, reduced chemical,and spring cover crop treatments) affect insect pest populations, arthropod predator communities, weedassemblages, and soybean yield and profitability. Soybean aphids exceeded economic thresholds in allthree years, and insecticides successfully suppressed these outbreaks in the two chemical treatments;aphids exceeded the economic injury level in the cover crop treatment in two of three study years. Beanleaf beetle (Cerotoma trifurcata Forster; Coleoptera: Chrysomelidae) populations were sub-economic inall treatments; insecticides targeting soybean aphid also reduced bean leaf beetles in the first year ofstudy when beetle populations were at their highest. Foliar-dwelling predator populations weresubstantially higher in the cover crop treatment than in the chemical treatments in all years of study;population declines in the latter treatments were strongly associated with insecticide applications tar-geting soybean aphids. Foliar predator populations did not rebound within the growing season afterinsecticides were applied. Soil predator populations were largely unaffected by treatment (except in2006, when they were more abundant in the cover crop treatment than in the chemical treatments).Weed communities varied among treatments and study years, with few consistent trends except that thechemically intensive treatment had lower weed densities than the other treatments. Although inputcosts of the cover crop and reduced chemical treatments were lower than the chemically intensivetreatment, the chemically intensive treatment was the most profitable of the three. Nevertheless, wecontend that the cover crops can be managed more efficiently in order to increase the profitability andcompetitiveness of this treatment while gaining the long-term benefits gleaned from conservingbiodiversity in our agroecosystems.

Published by Elsevier Ltd.

1. Introduction

The Northern Great Plains lies on the western edge of soybean(Gycine max [L.]) production in North America, and as such harborsa regionally adapted suite of insects and weeds. Soybeans in this

and surrounding regions were largely free from insect pests prior tothe introduction of the Asian soybean aphid (Aphis glycines Mat-sumura [Hemiptera: Aphididae]) (Ragsdale et al., 2004, 2011).Aphid populations rapidly increase during outbreak years (whichdo not occur every year), alter the physiology of soybean plants,reduce soybean yields, and potentially transmit soybean viruses(Clark and Perry, 2002; Macedo et al., 2003; Riedell and Catangui,2006; Beckendorf et al., 2008; Riedell et al., 2009). Bean leafbeetle (Cerotoma trifurcata Forster [Coleoptera: Chrysomelidae]) is

* Corresponding author. Tel.: þ1 605 693 5211; fax: þ1 605 693 5240.E-mail address: Jonathan.Lundgren@ars.usda.gov (J.G. Lundgren).

Contents lists available at SciVerse ScienceDirect

Crop Protection

journal homepage: www.elsevier .com/locate/cropro

0261-2194/$ e see front matter Published by Elsevier Ltd.http://dx.doi.org/10.1016/j.cropro.2012.08.005

Crop Protection 43 (2013) 104e118

This article is a U.S. government work, and is not subject to copyright in the United States.

another predominant, but sporadic, pest of soybeans in this region.Adults of this pest defoliate soybean plants and transmit bean podmottle virus (Mabry et al., 2003; Bradshaw et al., 2008; Byamukamaet al., 2011), and larvae consume root nodules of the soybean plantand disrupt its nitrogen dynamics (Lundgren and Riedell, 2008;Riedell et al., 2011). These two pests support a diverse and abun-dant natural enemy community throughout much of the soybean-producing region of North America (Toepfer et al., 2009; Ragsdaleet al., 2011), but this natural enemy community has been poorlydescribed in the Northern Great Plains (but see Seagraves andLundgren, 2012 for one report). Although soybeans are commonlyproduced in the Northern Great Plains (Nebraska, South Dakota andNorth Dakota were ranked 6th, 8th, and 9th in 2011 soybeanhectares harvested; National Agricultural Statistics Service, www.nass.usda.gov), there have been no peer-reviewed, comprehen-sive studies of insect communities in this crop after the invasion ofthe soybean aphid. This is particularly important in light of the factthat the introduction of soybean aphid has accompanied a dramaticincrease in insecticide use (Ragsdale et al., 2004; Fausti et al., inpress), and this undoubtedly has changed the dynamics of insectcommunities in this region substantially.

Pests are often a symptom of producing crops under mono-culture conditions, and efforts to increase the diversity of vegeta-tion within cropland often increase natural enemy populations andreduce pest intensity (Andow, 1991; Tillman et al., 2004; Broadet al., 2009; Lundgren and Fergen, 2010; Letourneau et al., 2011;Lundgren and Fergen, 2011; Koch et al., 2012). Sources of thisvegetation diversity may be in the form of low levels of weeds orthe use of cover crops or ground covers. In-season ground covers insoybeans can reduce soybean aphid populations, likely througha combination of altered plant quality for pest (soybean aphids andpotato leafhopper [Empoasca fabae Harris]) development andaugmented populations of natural enemies (Miklasiewicz andHammond, 2001; Schmidt et al., 2007). But these understoryground covers can also reduce soybean yields to such levels thatsome researchers concluded that their use is unrealistic forproducers (Schmidt et al., 2007). Cover crops do not necessarilyhave these negative effects on soybean yields (Davis, 2010; Smithet al., 2011). Weed populations provide another source of vegeta-tion diversity that is used by natural enemies of soybean aphids(Griffen and Yeargan, 2002; Lundgren et al., 2009b). The effects ofweed presence on soybean aphids has not been well studied, but islikely going to be more important as glyphosate-resistant pop-ulations of weeds expand their ranges (Lundgren et al., 2009a;Heap, 2012; Mortensen et al., 2012). Because input costs varysubstantially depending on pest management approaches, andinsect management decisions are linked to other aspects of cropproduction, the benefits of vegetation diversity on pest manage-ment of soybean insects can only be verified using a systems-levelapproach that incorporates the relative economic costs and benefitsof different best-practice management philosophies.

We used a systems-level approach to examine the relativeeffects of three soybean production systems on insect and weedcommunities and soybean profitability at one location over threeyears in the Northern Great Plains. Specifically, we compared theeffects of a cover crop-based systemdesigned tominimize chemicalinputs and a system intended to reduce herbicide inputs (andharbor low levels of weeds) with a system reliant on high chemicalusage typical of our region. These treatments varied across yearsand even among plots depending on the prevailing conditionsobserved at the sites, but were unified through the underlyingphilosophies that defined the treatments. Although systems-levelprojects such as these lack the ability to identify clear mecha-nisms for how specific results are produced within a particularsystem, this approach provides a ‘real world’ picture of how these

systems will affect key crop production characteristics for end-users. In addition to examining the relative costs and benefits ofthese different systems, this study represents the most compre-hensive recent survey of insect communities of soybeans in theNorthern Great Plains published in the peer-reviewed literature.

2. Methods

2.1. Experimental field sites

Research was conducted in 2005e2007 on the Eastern SouthDakota Soil and Water Research Farm, near Brookings, SD (latitude,longitude: 44.348, �96.811). Experimental plots were24.4 � 30.5 m in 2005, 12.1 �18.3 m in 2006, 18.3� 18.3 m in 2007,and the three treatments were arranged in a RCB design with threereplicates in 2005 (nine total plots) and four replicates in 2006 and2007 (12 total plots). At least 6 m margins separated the plots;these margins were spring-planted to winter wheat (Triticum aes-tivum L.) of mixed varieties at 111 kg/ha. The three treatmentsexamined were 1) chemically intensive management, 2) reducedchemical management, and 3) cover crop-based management(each is described more extensively below). Experiments wereembedded in a split, 12.5-ha field; the two halves were rotatedbetween corn and soybeans so that half the field was soybeans eachyear. Soybeans were planted at approximately 370,000 viable seedsper ha on 6-June 2005, 9-June 2006, and 7-June 2007 (using 0.76 mrow spacing). Throughout the experiment, the chemically intensiveand reduced chemical systems were planted to soybean variety91M40 (a glyphosate tolerant variety; Pioneer HiBred, Johnston,IA). The cover crops systemwas initially planted to 91M40 in 2005,but was switched to 91M10 soybeans in 2006 & 2007; 91M10 is notglyphosate tolerant, and thus fits within the philosophy of thismanagement system better than the 91M40. The field was underconventional tillage (fall and spring chisel plow), and received224 kg/ha of 14e36e13 (NePeK; ammonium phosphate) in 2005.In 2006 and 2007, no fertilizer or inoculant was applied. Plots wereharvested by combine in mid- to late-October, when grainmoisturereached approximately 13% as measured using a grain moisturetester (DICKEY john Corporation, Auburn, IL). Grain yields werecalculated, and grain quality was assessed (% dry matter, oil, andprotein) using a near infrared spectrometer (Foss North America,Eden Prairie, MN). For these grain metrics, three subsamples weretaken per plot. Two of the 2007 plots could not be harvested due tohighweed pressure (one in the cover-crop based treatment and onein the reduced chemical treatment); yields were considered as zerofor these plots, and the data from these plots were excluded fromthe nutrient analyses. Average maximum seasonal temperatureswere 26.11 (2005), 25.00 (2006) and 26.11 �C (2007), and totalprecipitation from June 1 to September 30 was 52.37 (2005), 37.13(2006) and 27.41 (2007) cm.

2.2. Chemical intensive system

This study system is meant to represent conventional produc-tion practices of our region. Specifically, pre-emergent herbicideand glyphosate are used to manage weed populations, and insec-ticides are applied as soybean aphids exceed economic thresholds.At planting in 2005, glyphosate (RoundUp, Monsanto Company, St.Louis, MO) was applied at 4.7 L (1.66 kg ai) per ha tank-mixed withthe pre-emergent herbicide S-metolachlor (Dual Magnum�, Syn-genta, Greensboro, NC) at 2.3 L (2.1 kg ai) per ha. In 2006, the samerate of glyphosate was applied, but the S-metolachlor (Dual IIMagnum, Syngenta) pre-emergent herbicide was applied at 1.2 L(1.1 kg ai) per ha. In 2007, 2.3 L (0.83 kg ai) per ha of glyphosatewas applied at planting without a pre-emergent herbicide. An

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118 105

additional application of glyphosate (1.6 L [0.6 kg ai] per ha) wasapplied when soybeans were at the V3 stage (July 1, 2005, July 18,2006, July 6, 2007).

Insecticides were applied to this treatment when soybean aphidpopulations exceeded the economic threshold of 250 aphids perplant (Ragsdale et al., 2007). This occurred on August 12, 2005 andAugust 13, 2007. At this point, lambda cyhalothrin (Warrior�,Syngenta Corporation) was applied at 0.14 L [0.011 kg ai] per ha. In2006, esfenvalerate (Asana� XL, DuPont, Wilmington, DE) wasapplied at 0.58 L (0.04 kg ai) per ha when aphids initially exceededthe economic threshold (August 7, 2006). Aphids exceeded theeconomic threshold 7 d following this spray, and 0.16 L (0.012 kg ai)of lambda cyhalothrin per ha was applied on August 18, 2006.

2.3. Reduced chemical system

In this treatment, we reduced the herbicide inputs by onlymanaging weed populations at planting, with an additional inputas needed up until the V3 stage of the soybeans (this only occurredin 2005). In all experimental years, glyphosate alone was applied at4.67 L (1.66 kg ai) per ha at planting. In 2005, all plots received anadditional application of glyphosate (1.61 L [0.6 kg ai] per ha) onJuly 1. No additional herbicides were applied to these plots in 2006or 2007. Insecticides were applied when aphid populationsexceeded economic thresholds. No insecticides were applied to thistreatment in 2005 as the economic threshold for aphids was onlyexceeded after the soybean canopy closed, which prohibited tractorapplication of insecticides. Esfenvalerate (Asana XL, Dupont) wasapplied at 0.58 L (0.04 kg ai) per ha to two of the plots on August 7,2006 and lambda cyhalothrin (Warrior, Syngenta) was applied at0.16 L (0.013 kg ai) per ha to the remaining two plots on August 15,2006. Lambda cyhalothrin (0.14 L [0.011 kg ai] per ha) was appliedto all plots on August 13, 2007.

2.4. Cover crop-based system

In this system, a winter cereal cover crop was planted early inthe spring prior to planting the soybeans, andwas eventually killed.This cover crop replaced additional herbicides and all insecticidesin this management system. In 2005 and 2007, winter oat (Avenasativa L.) var. Jerry (IL Foundation Seeds, Champaign, IL) wasplanted at 77.3 kg per ha on April 8, 2005 and May 14, 2007. In2006, winter rye (Secale cereale L.; Millborn Seeds, Brookings, SD)was spring-planted at 127 kg per ha on May 18. The oat crop wasmowed to about 4.5 cm tall on June 2, 2005 and 7.5 cm on June 16,2005 using a flail mower (soybeans were shorter than this cut line).Glyphosate (1.6 L [0.6 kg ai] per ha) was applied to this treatment onJuly 1, 2005. The rye was mowed using a flail mower on July 6,2006; the rye plants were heading and about 1 m tall. The soybeanswere undamaged by the mowing. Herbicide was not applied to thistreatment in 2006, and thus the rye persisted throughout theseason. In 2005 and 2006, the cover crop was competitive withsoybean. Therefore, in 2007 the oat crop was treated with anapplication of glyphosate (2.33 L [0.83 kg ai] per ha) on June 4whenthe soybean crop was planted. Any subsequent benefits were theresult of the cover crop residue.

3. Insect monitoring

3.1. Aphid sampling

Aphids were counted visually on whole plants in the fieldthroughout each growing season. The numbers of sampling datesand plants sampled per date changed according to resources andaphid population sizes. The subsampling system developed by Yoo

and O’Neil (2009) was applied in 2006 and 2007 to aphid pop-ulations that exceeded the economic injury level (EIL; approxi-mately 694 � 95 aphids per plant [mean � 95% CI], but note thatthis EIL would be lower when soybean prices exceed $238 permetric ton; Ragsdale et al., 2007). The economic threshold (ET), orthe aphid population that warrants treatment to avoid exceedingthe EIL, is 250 aphids per plant. The scheme involved sampling onlythe trifoliates on the 1st, 5th, and 9th nodes (from the top), andmultiplying the resulting number by 3.76 to attain total aphidpopulations on the plant. In 2005, aphids were counted on 12 datesfrom June 24 to September 8, with sample numbers diminishing asplant size aphid populations increased from 100 plants per plot inJune to 6 plants per plot at the end of the season. In 2006, plotswere sampled 12 times between June 29 and September 8 with 20plants per date up to August 11, and 15 plants thereafter. During2007, 13 samples were taken from June 19 to September 6, with 20plants per plot on all dates.

3.2. Tile traps for aphid colonization

Green tile traps monitored the influx of winged soybean aphidsinto plots (Hodgson et al., 2005). Individual traps consisted ofa collection reservoir (3 cm diameter � 5 cm height) attached toa 1.5 m tall metal rod. The collection reservoir was composed ofa square green tile (11 �11 cm; H & R Johnson, Stoke-on-Trent, UK)that mimicked the reflectance spectrum of soybean leaves (Irwinand Goodman, 1981). The tile was secured to the bottom half ofa clear plastic sandwich box (11 � 11 � 3 cm), which was glued tothe lid of a cylindrical plastic canister. One trap was centered ineach plot by inserting the metal rod into the ground betweensoybean rows, and trap height was adjusted equal to the plantcanopy height throughout the trapping period (26 June to 27 July2006, 22 June to 3 August 2007). Trapping was terminated whenalatoid offspring of soybean aphids began appearing on soybeanplants within the plots. Each trap reservoir receivedw200ml of 1:1solution of water and ethylene glycol to retain colonizing aphids.Trap contents were emptied every 3e4 d into a jar, and freshsolution was added to the trap reservoir. Alate soybean aphidswere identified and tallied for each sample date. Tallies fromindividual traps were summed across sampling dates prior todata analysis.

3.3. Sweep samples

Soybean foliage was swept with a 38-cm diameter net period-ically throughout each season from late June (prior to aphid colo-nization) until the late R6 growth stage (wmid September). In2005, 75 sweeps per plot were collected on each of 10 sample dates(June 24, July 1, 8, 15, 22, August 3, 16, 26, September 2, and 14). In2006, sweep samples were collected on each of 12 sample days(June 22, 30, July 7, 14, 20, 28, August 4, 14, 21, 29, September 6, and13, 2006); twenty sweeps per plot were collected before July 7, and60 per plot were collected on June 7 and afterward. In 2007, 75sweeps per plot were collected on each of 11 sample days (June 26,July 5, 12, 19, 26, August 3, 9, 17, 24, 31, and September 6, 2007). Thenumber of each species per sweep was calculated from thesesamples.

In all three seasons, the number of Orius insidiosus (Say)(Hemiptera: Anthocoridae), Coccinellidae adults (identified tospecies), larvae of Coccinellidae, Chrysopidae adults (Chrysoperlaspp.), and larvae of Chrysopidae were counted. Additionally, thenumbers of bean leaf beetles were recorded in all three years. Thenumber of Geocoris sp. (Hemiptera: Geocoridae) were counted insamples collected after July 14, 2006 (inclusive). Hemerobiidae(Neuroptera) and Nabidae (Hemiptera; likely Nabis americoferus

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118106

Carayon) were added in 2006 and 2007. Spiders (Araneae) andharvestmen (Opiliones: Phalangiidae) were frequently found, butwere not recorded from the sweeps.

3.4. Pitfalls

Five pitfall traps were established in each plot in a central-ized � pattern. The four lateral traps were 9.0 m from the centraltrap in 2005 (6.8 m in 2006 & 2007). Each trap consisted of a hole inthe ground lined with a PVC tube (10 cm diam., approximately20 cm deep). To activate the traps, a glass jar (8 cm diam., 12 cmdeep) filled with 200 ml of ethylene glycol was placed into eachtube; a funnel (openings of 10 and 2 cm) was inserted on top of thejar, and the entire assembly was loosely covered with a plywoodboard (0.3 � 0.3 m). Pitfall series were activated for 5 d at a time.Over the duration of the experiment, seven pitfall series weredeployed in 2005 (collected on May 9, May 23, June 22, July 11,August 1, August 19, and September 16), six series were deployed in2006 (collected on May 30, June 28, July 11, July 25, August 8,August 25), and five series were deployed in 2007 (collected onJune 13, July 3, July 24, August 13, and September 4). After thecollection period, samples were returned to the lab, and specimenswere washed, cleaned, and identified to the lowest taxonomicposition possible. In 2005, spiders, carabid larvae, chrysopid larvae,predatory hemipterans, lampyrid larvae, and centipedes were notquantified, but they were in subsequent years.

3.5. Quadrat samples

Actual densities of surface-dwelling predators were monitoredusing quadrat samples (Lundgren et al., 2006; Lundgren andFergen, 2010). A quadrat made of sheet metal (0.5 � 0.5 m; 15 cmtall) was randomly placed within a plot for each sample observa-tion, and the insects found within the top 1 cm of soil were aspi-rated. In 2005, we collected three samples per plot on each of ninesample dates (May 6, 9, 23, June 22, July 11, 22, August 1, 19, andSeptember 16). In 2006, two samples were collected from each ploton five sample dates (June 21, July 11, July 25, August 8, and 23).Nocturnal samples were collected (between 22:00 and 2:00) fromeach plot on July 11e12 (three per plot) and August 8e9 (two perplot), 2006. In 2007, two samples were collected from each plot onfive sample dates (June 12, 27, July 11, 26, and August 8). The insectscollected were curated and identified to as low a taxonomic level aspossible. The mean number of predators per m2 on the soil surfaceof each plot was extrapolated from these measurements.

3.6. Weed populations

Weed species diversity and relative abundance by species werecounted along two cross-plot diagonal line transects every 0.5 m.Counts were taken four times in 2005 (at planting, V2, V3, and aftercanopy closure) and three times in 2006 and 2007 (at planting, V3growth stage, and after canopy closure). The sampling dates wereJune 2, June 27, July 11, and 28, 2005; June 6, July 6, and August 31,2006; and June 7, July 3, and July 18, 2007.

3.7. Data analysis

In all analyses, data were analyzed separately for each studyyear, and data at least approximately conformed to the assumptionsof ANOVA. The number of aphids per plant, bean leaf beetles persweep, predators (per sweep, pitfall trap or m2 quadrat), and weedsper plot were compared among treatments using independent,random, repeated-measures (rm) ANOVAs; all independent vari-ables were considered random effects, and treatment and sample

dates were considered fixed effects. The seasonal abundance ofalate aphids captured per tile trap, individual predator and weedtaxa collected, and the number of weed species collected werecompared using independent, random one-way ANOVAs. Statisti-cally different means were separated using LSD means separationtests. All statistics were performed using SYSTAT 11 (SYSTAT Soft-ware, Inc., Chicago, IL, 60606).

Data on production input and application methods were used togenerate the cost of production estimates. Actual cost data werecollected from USDA reports, SDSU extension personnel, andprivate sector sources. Cost data sources are listed in Appendix 1.An Ordinary Least Squares (OLS) regression analysis was conductedto gain greater insight on treatment and exogenous factors affectingyield variability. Our unit of measure is kg per ha (Yt). The data setcontains 92 usable observations. We created three bi-variatedummy variables for both year and production method (e.g., oneif year is 2005, zero otherwise or one if cover crop treatment, zerootherwise). The dummy variables are Yr05, Yr07, RECHEM (reducedchemical treatment), and COVER (cover crop treatment). Thus, thebase variables are 2006 and traditional chemical treatment. Toaccount for the potential effect of oat being used as the cover cropin years 2005 and 2007, but rye being used in 2006 we createda dummy variable to account for this change (RYE). The regressionequation is defined as:

Yt ¼ aþ b1RECHEMt þ b2COVERt þ b3YR05t þ b4YR07t

þ b5RYEt þ Ut ;

where t denotes year, a the intercept term, b the regression coef-ficient, and Ut the regression error term. Regression estimates areprovided in Table 6, and diagnostics were conducted to determine ifany of the OLS assumptions were violated. It was determined thatthe regression residuals displayed a heterogeneous pattern thatrequired the application of White’s correction procedure for het-eroscedasticity (White, 1980). Heteroscedasticity refers to a viola-tion of the OLS assumption of the OLS error term having a constantvariance. Failure to correct for this problem results in unreliablestandard error estimates.

4. Results

4.1. Pest populations

Aphid densities varied among management systems and dateseach year (P< 0.001 for all comparisons; Fig. 1), and this is reflectedin how quickly aphid populations reached the economic threshold(ET) and economic injury levels (EIL) in these plots. In 2005, the ETwas only exceeded on 12-August in the chemically intensive plots,and their numbers were reduced after application of lambda-cyhalothrin. The ET was exceeded on 18-August in the reducedand cover crop management systems. No insecticide was applied toeither treatment, in the case of reduced chemical management dueto canopy closure and in the case of cover crop due to no insecticidedecision. The EIL (674 aphids per plant) was exceeded in all plots ofthe reduced chemical and cover crop treatments on at least onesample date after 25-August. Aphid populations escalated in thecover crop treatment until sampling ceased on 8-September (theEIL was exceeded in all plots after 1-September). In 2006, the ETwas exceeded in the chemically intensive and reduced chemicalplots on about Aug 6, and remained above the ET until lambda-cyhalothrin was applied on 15-August; mean aphid abundancenever exceeded the EIL in either of these treatments, and aphidsremained well below the ET throughout the season in the covercrop treatment. In 2007, all three management types exceeded the

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118 107

EIL for a single date (August 10). Insecticides in the chemicallyintensive and reduced chemical management systems reduced theaphid population. In the cover crop treatment, aphids remainedabove the EIL from August 10 to 30. There were no differencesamong treatments in the total number of alate aphids recoveredfrom the tile traps in each plot in either year of study (2006 F2,9 ¼ 0.49, P ¼ 0.63; 2007: F2, 9 ¼ 0.47, P ¼ 0.64). Mean (SEM)cumulative alate aphids (pooled across treatments) recovered per

plot over the seasons were 36.92 � 2.42 (2006) and 42.92 � 4.82(2007). Alate aphids only became abundant (>2 per trap) after 20-July, 2006 and 24-July, 2007.

There were fewer bean leaf beetles in the chemically intensivetreatment in 2005 than in the other treatments (treatment: F2,6 ¼ 5.27, P ¼ 0.048; date: F9, 54 ¼ 25.09, P < 0.001; interaction: F18,54 ¼ 4.13, P < 0.001) (Table 1). Bean leaf beetle populations weresimilar among treatments in the subsequent two seasons (Table 1)(2006: treatment: F2, 9 ¼ 0.65, P ¼ 0.55; date: F7, 63 ¼ 8.35,P< 0.001; interaction: F14, 63 ¼ 3.51, P< 0.001; 2007: treatment: F2,9 ¼ 0.69, P ¼ 0.53; date: F10, 90 ¼ 32.46, P < 0.001; interaction: F20,90 ¼ 1.70, P < 0.001). Beetle populations varied over the season,displaying two population peaks over the season as has beenpreviously reported in our region (Hammack et al., 2010; Riedellet al., 2011). There were significant interactions between treat-ment and sample date in all three years. Population trends over theseasons were similar among treatments in 2005 and 2007 (e.g., allhad peaks and troughs on the same sample dates), but they variedin the degree of change over the season (peaks and troughs weremore severe in some treatments rather than others). This was alsotrue in 2006, but the cover crop treatment had a large peak in beanleaf beetle populations toward the end of the season that was notpresent in the other treatments. Seasonal mean � SEM numbers ofbeetles collected per sweep in each treatment year are presented inTable 1.

4.2. Natural enemy populations

We discovered an abundant and diverse natural enemycommunity present in South Dakota soybeans. In sum, 15,378predators were collected in sweep samples, 18,092 in pitfallsamples, and 6688 in quadrat samples over this 3 yr study. At least17 species (representing five families) were collected in sweepsamples, 65 taxa (representing 13 families) in pitfall samples, and44 taxa (representing 17 families) in quadrat samples (Tables 1e3).Based on our density samples, we estimate that there are about277,170 � 30,283, 339,300 � 84,100, and 594,800 � 133,300predators per ha on the soil surface in the chemical intensive,reduced chemical, and cover crop management systems, respec-tively. These numbers are within the range of previous estimates ofnatural enemies in cropland (Lundgren et al., 2006; Lundgren andFergen, 2010).

The foliar dwelling natural enemy community was reduced bythe insecticide applications, but the soil-surface community wasunchanged due to treatment. Sweep samples revealed that foliardwelling arthropods in all three growing seasons were affected bymanagement and sampling date (Table 1, Fig. 2) (2005: treatment:F2, 6¼ 20.78, P¼ 0.002; date: F9, 54¼ 9.46, P< 0.001; interaction: F18,54¼ 3.61, P< 0.001; 2006: treatment: F2, 9¼103.37, P< 0.001; date:F11, 99¼ 32.08, P< 0.001; interaction: F22, 99¼ 9.85, P< 0.001; 2007:treatment: F2, 9 ¼ 239.08, P< 0.001; date: F10, 90 ¼ 64.52, P< 0.001;interaction: F20, 90 ¼ 32.68, P < 0.001). Although sample dateaffected the number of foliar-dwelling natural enemies collected inall three treatments, the significant interaction suggests that thiseffect was not consistent among treatments (treatment by dateinteraction P < 0.001 each year). This is evident from Fig. 2, wherenatural enemies in the chemically treated systems were reducedafter application relative to those in unsprayed treatments. Addi-tional predators that were rarely collected in the sweep samplesincluded several Coccinellidae species: Cyclonedamunda (a total of 9specimens collected), Hippodamia variegata (Goeze) (n ¼ 1), Bra-chyacantha ursina (Fabricius) (n ¼ 8) and B. albifrons (Say) (n ¼ 1).Natural enemy activity and density on the soil surface was unaf-fected by treatment, and consistently increased over thefield season(Tables 2 and 3) (Pitfall samples: date P < 0.001 each year; Quadrat

Fig. 1. Soybean aphid population phenology in soybeans produced under threemanagement regimens. Data points represent mean (SEM) aphid numbers per plantper sample plot (n ¼ 3 plots in 2005, and 4 in 2006 & 2007). Short-dashed linesindicate the economic injury level and long-dashed lines indicate the economicthreshold for this pest. Arrows indicate dates when insecticides were applied to thechemically intensive (all years) and reduced chemical (only in years 2 & 3) treatments.Statistical outputs were: 2005 management system: F2, 6 ¼ 14.42, P ¼ 0.008, time: F11,66 ¼ 17.14, P < 0.001, treatment � time: F22, 66 ¼ 6.34, P < 0.001. 2006 managementsystem: F2, 9 ¼ 17.70, P < 0.001, time: F10, 90 ¼ 12.90, P < 0.001, treatment � time: F20,90 ¼ 4.61, P < 0.001. 2007 management system: F2, 9 ¼ 13.76, P ¼ 0.002, time: F11,99 ¼ 15.92, P < 0.001, treatment � time: F22, 99 ¼ 10.58, P < 0.001.

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118108

samples:

date

P<

0.002

eachyear).A

nexcep

tionwas

thenatu

ralen

emyden

sityin

2006;

here,

natu

ralen

emies

were

much

more

abundan

tin

thecover

cropped

treatmen

tthan

intheoth

ertw

oman

agemen

tsystem

s(Table

3).Addition

alpred

atorscollected

rarelyin

thepitfalltrap

sinclu

ded

:Agonum

cupripenne(Say)(a

totalof

4sp

ecimen

scollected

),Amara

apricaria(Payku

ll)(n

¼1),

A.carinata

(LeCon

te)(n

¼6),

A.littoralis

Man

nerh

eim(n

¼1),

Amphasia

sericeus(T.

W.Harris)

(n¼

4),Anisodactylus

ovularis(Casey)

(n¼

1),A.sanctaecrucis

(Fabricius)

(n¼

3),anuniden

tified

Bem

bidiin

e(n¼

1),Bradycellusneglectus

(LeCon

te)(n¼1),Calathus

Fig.2.Totalinsect

predatorabundance

persw

eepin

soybeansproduced

underthree

managem

entregim

ens.Data

pointsrepresent

mean

(SEM)num

berof

predators(taxa

arelisted

inTable

1)collected

persw

eepper

plot(n¼

3plots

in2005,and

4in

2006&

2007).Arrow

sindicate

dateswhen

insecticideswere

appliedto

thechem

icallyintensive

(allyears)and

reducedchem

ical(only

inyears

2&

3)treatm

ents.

Table 1Bean leaf beetles and major foliar-dwelling insect predators collected in soybeans under three management regimens. Insects were collected using sweep nets. Data represent the mean (SEM) number of insects per 100 sweepscollected per sample date per plot (n¼ 3 plots per treatment in 2005 and n¼ 4 in 2006 & 2007). Significant treatment effects (a¼ 0.05) are indicated in bold italics. Within years and parameters, treatment means without lettersindicate that the F-test was not significant and no pairwise testing was done. Where letters are present after means, the treatment F-test was significant (a ¼ 0.05), and means followed by the same letter are not significantlydifferent (LSD multiple comparison test). Additional taxa that were infrequently collected are listed in the results section.

2005 2006 2007

Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical

PestsCerotoma trifurcata 6.18 � 1.96a 26.80 � 4.55b 22.84 � 6.68b 1.60 � 0.26 2.36 � 1.04 1.42 � 0.07 3.03 � 0.65 2.33 � 0.36 2.24 � 0.51Natural enemiesNeuroptera: ChrysopidaeChrysoperla sp. adult 0.98 � 0.27a 4.49 � 0.54b 1.78 � 0.27a 2.12 � 0.24 2.08 � 0.43 2.33 � 0.34 2.42 � 0.35a 11.09 � 2.98b 1.55 � 0.58aChrysoperla sp. larva 0.27 � 0.15a 1.78 � 0.27b 0.89 � 0.09a 0.69 � 0.13 1.18 � 0.04 1.25 � 0.32 1.76 � 0.42 3.58 � 0.83 1.48 � 0.23Coleoptera: CoccinellidaeCoccinellid larvae 0.31 � 0.09a 22.68 � 7.92b 10.68 � 0.87ab 4.51 � 1.56 2.78 � 0.85 4.48 � 0.86 7.06 � 1.85a 15.94 � 1.89b 5.55 � 0.42aCoccinella septempunctata (L.) 0.09 � 0.04a 0.75 � 0.16b 0.09 � 0.09a 0.21 � 0.09 0.35 � 0.13 0.28 � 0.15 0.73 � 0.16a 2.24 � 0.40b 0.58 � 0.12aColeomegilla maculata De Geer 0.09 � 0.04 2.04 � 1.06 1.55 � 0.23 0.28 � 0.10 0.87 � 0.31 0.21 � 0.09 0.09 � 0.03a 2.24 � 0.38b 0.52 � 0.21aCycloneda munda (Say) 0 0 0 0 0 0.03 � 0.03 0a 0.27 � 0.10b 0aHarmonia axyridis (Pallas) 1.60 � 0.08 16.72 � 7.13 9.79 � 2.90 0.90 � 0.26 3.51 � 1.02 2.29 � 0.46 1.97 � 0.26a 24.00 � 3.03b 1.61 � 0.52aHippodamia convergens Guérin-Méneville 0.4 � 0.34a 4.71 � 0.78b 0.89 � 0.56a 3.65 � 0.26 5.76 � 0.70 6.07 � 1.13 0.45 � 0.06a 2.85 � 0.63b 0.18 � 0.10aHippodamia parenthesis (Say) 0.04 � 0.04a 0.57 � 0.16b 0.22 � 0.04a 1.28 � 0.15 3.92 � 1.34 1.42 � 0.35 0.18 � 0.03 0.27 � 0.10 0.09 � 0.06Hippodamia tredecempunctata (Say) 0a 0.31 � 0.04b 0.04 � 0.04a 0.24 � 0.07 0.28 � 0.11 0.21 � 0.07 0.03 � 0.03 0.12 � 0.07 0.09 � 0.03Scymnus rubricaudus (Casey) 0.27 � 0a 16.40 � 3.32b 1.96 � 0.66a 0.31 � 0.07a 7.88 � 1.48b 0.49 � 0.12a 0.15 � 0.06 0.33 � 0.13 0.09 � 0.09Hemiptera: Geocoridae, Nabidae, and AnthocoridaeGeocoris sp. NA NA NA 1.85 � 0.52 12.78 � 1.60 2.41 � 0.62 1.18 � 0.34 1.36 � 0.09 1.06 � 0.37Nabis americoferus Carayon NA NA NA 14.72 � 0.67a 46.08 � 4.35b 20.42 � 1.57a 14.88 � 1.06a 26.52 � 1.74b 13.36 � 0.98aOrius insidiosus (Say) 3.73 � 0.15c 12.27 � 0.68a 7.81 � 0.79b 6.08 � 0.70a 15.63 � 2.40b 7.15 � 0.52a 11.51 � 0.68a 35.48 � 4.66b 14.70 � 0.87aTotal predators 7.96 � 0.25a 82.73 � 13.16b 35.39 � 5.77a 36.42 � 2.79a 100.45 � 4.25c 48.47 � 2.78b 40.45 � 2.56a 100.23 � 2.97b 39.30 � 0.96a

J.G.Lundgren

etal./

CropProtection

43(2013)

104e118

109

Table 2The activity densities of major predatory arthropods collected from soybeans managed under three different regimens. Activity densities were estimated using pitfall traps. Data represents the mean (SEM) number of individualtaxa collected per trap, per sample period, per plot (n¼ 3 plots per treatment in 2005 and n¼ 4 in 2006 & 2007). Significant treatment effects (a¼ 0.05) are indicated in bold italics. Within years and parameters, treatment meanswithout letters indicate that the F-test was not significant and no pairwise testing was done. Where letters are present after means, the treatment F-test was significant (a ¼ 0.05), and means followed by the same letter are notsignificantly different (LSD multiple comparison test). Additional taxa that were infrequently collected are listed in the results section.

2005 (seven sample periods) 2006 (six sample dates) 2007 (five sample dates)

Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical

AraneaeAraneae spp. NA NA NA 3.88 � 0.44 10.08 � 0.73 14.25 � 10.64 2.35 � 0.67 3.75 � 0.73 3.20 � 0.85Lycosidae spp. NA NA NA 0.54 � 0.17 0.71 � 0.20 1.29 � 0.28 0.70 � 0.26 0.50 � 0.17 0.90 � 0.17ChilopodaChilopod sp. NA NA NA 0.50 � 0.22 0.29 � 0.10 0.58 � 0.26 0.40 � 0.24 0.55 � 0.36 0.90 � 0.37Coleoptera: CarabidaeColeoptera larvaea NA NA NA 0.25 � 0.11 0.13 � 0.13 0.25 � 0.08 0.30 � 0.13 0.25 � 0.10 0.35 � 0.10Agonum placidum (Say) 0.81 � 0.22 0.19 � 0.06 0.29 � 0.11 1.58 � 0.56 2.67 � 0.82 1.38 � 0.64 8.35 � 0.69 9.70 � 1.73 11.45 � 2.15Bembidion quadrimaculatum oppositum (Say) 0.05 � 0.05 0.05 � 0.05 0.05 � 0.05 0.29 � 0.13 0.46 � 0.08 0.08 � 0.08 0.05 � 0.05 0 0.15 � 0.10Bembidion rapidum (Leconte) 0 0.05 � 0.05 0 0.17 � 0.07 0.21 � 0.16 0 0.10 � 0.10 0.20 � 0.08 0.10 � 0.10Brachinus ovipennis LeConte 0.29 � 0.08 1.76 � 0.55 0.67 � 0.39 0.54 � 0.86 0.08 � 0.7 0.33 � 0.20 0.75 � 0.13 1.00 � 0.18 0.95 � 0.17Cyclotrachelus alternans (Casey) 5.81 � 1.40 13.43 � 5.08 9.10 � 2.20 6.92 � 1.34 4.79 � 0.66 9.00 � 2.22 8.80 � 1.00 8.80 � 1.28 9.10 � 0.44Elaphropus sp. 0.10 � 0.10 0.05 � 0.05 0.29 � 0.14 1.13 � 0.57a 2.79 � 0.26b 1.08 � 0.49a 0.65 � 0.32 1.70 � 0.44 2.25 � 0.88Harpalus pensylvanicus (DeGeer) 0 0.05 � 0.05 0.10 � 0.10 0.71 � 0.28 0.13 � 0.08 0.33 � 0.15 0.15 � 0.10 0.40 � 0.14 0.40 � 0.14Poecilus chalcites (Say) 3.33 � 1.93 3.24 � 0.34 7.05 � 0.10 0.50 � 0.30 0.21 � 0.04 0.67 � 0.32 2.00 � 0.76 2.50 � 0.56 1.90 � 0.49Poecilus lucublandus lucublandus (Say) 2.71 � 1.30 3.14 � 1.09 2.33 � 0.88 2.71 � 1.13 1.25 � 0.28 3.04 � 0.49 0.75 � 0.30 1.15 � 0.28 0.7 � 0.17Pterostichus permundus (Say) 2.62 � 0.52 2.05 � 0.37 1.48 � 0.70 1.58 � 0.26 1.71 � 0.56 1.58 � 0.43 9.15 � 1.56 8.95 � 1.30 6.10 � 0.70Scarites quadriceps Chaudoir 0.52 � 0.17b 2.10 � 0.64ab 2.90 � 0.62a 1.04 � 0.23 0.79 � 0.10 0.79 � 0.21 2.10 � 0.35 2.30 � 0.47 1.80 � 0.85Coleoptera: CoccinellidaeCoccinellid spp. 0.14 � 0.08 0.57 � 0.44 0 0.29 � 0.20a 1.25 � 0.38b 0.29 � 0.13a 0.05 � 0.05 0.10 � 0.10 0.10 � 0.10Coccinellid larvae 0 0.05 � 0.05 0.10 � 0.10 0.38 � 0.10 4.71 � 2.58 0.33 � 0.12 0 0.10 � 0.06 0.05 � 0.05Coleoptera: StaphylinidaeStaphylinid spp. 6.95 � 2.02 9.05 � 2.19 6.76 � 1.53 10.08 � 0.55 10.29 � 1.38 8.58 � 2.08 1.95 � 0.30 2.95 � 1.48 1.85 � 0.96Hymenoptera: FormicidaeFormicid spp. 0.67 � 0.21 0.52 � 0.33 0.10 � 0.05 14.00 � 8.83 20.04 � 5.56 17.71 � 10.43 14.50 � 5.13 8.30 � 2.14 19.40 � 2.05Opiliones: PhalangiidaePhalangium opilio L. 28.19 � 0.80 22.38 � 1.80 27.86 � 1.93 20.50 � 1.65 12.46 � 3.60 15.29 � 1.80 21.85 � 3.71 18.05 � 1.99 23.00 � 1.61Orthoptera: GryllidaeAllonemobius sp. 1.86 � 0.66 2.10 � 0.56 4.09 � 1.29 10.25 � 1.91 8.00 � 1.24 7.75 � 1.05 21.80 � 5.25 19.85 � 4.14 23.35 � 6.21Gryllus pennsylvanicus Burmeister 8.48 � 1.39 3.95 � 0.27 17.43 � 5.36 6.79 � 1.79 7.50 � 0.78 6.75 � 1.86 16.70 � 4.86 14.45 � 3.28 17.25 � 4.95Total predators 63.62 � 11.75 65.38 � 6.48 81.33 � 5.48 88.83 � 9.94 93.25 � 8.52 95.83 � 10.25 114.45 � 17.88 108.20 � 9.12 127.60 � 7.80

a Although these larvae were not identified to family, the vast majority belonged to unknown Carabidae species.

J.G.Lundgren

etal./

CropProtection

43(2013)

104e118

110

Table 3The density (per m2) of major predatory arthropods collected on the soil surface in soybean fields managed under three regimens. Densities were obtained by hand collections of specimens from metal quadrats. Data representthe mean (SEM) number of taxa collected per sample date per plot (n ¼ 3 plots per treatment in 2005 and n ¼ 4 in 2006 & 2007). Significant treatment effects (a¼ 0.05) are indicated in bold italics. Within years and parameters,treatment means without letters indicate that the F-test was not significant and no pairwise testing was done. Where letters are present after means, the treatment F-test was significant (a ¼ 0.05), and means followed by thesame letter are not significantly different (LSD multiple comparison test). Additional taxa that were infrequently collected are listed in the results section.

2005 2006 2007

Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical Chemical intensive Cover crop Reduced chemical

AraneaeAraneae spp. 6.86 � 1.65 13.19 � 3.94 7.21 � 1.51 2.55 � 0.33 13.64 � 1.41 21.24 � 17.78 6.10 � 0.30 4.40 � 0.54 4.30 � 01.33Lycosid spp. 0.69 � 0.18 1.73 � 0.39 1.19 � 0.23 0.21 � 0.14 0.24 � 0.12 0.14 � 0.08 0 0 0ChilopodaChilopod sp. 0.05 � 0.05 0.05 � 0.05 0.30 � 0.15 0.64 � 0.21 0.71 � 0.30 0.36 � 0.14 0.30 � 0.19 0.20 � 0.12 0.50 � 0.25Coleoptera: CarabidaeCarabid larvae 0.40 � 0.20 0.44 � 0.23 0.64 � 0.27 0.36 � 0.07a 3.38 � 1.06b 1.21 � 0.21a 0.10 � 0.10 0.10 � 0.12 0.10 � 0.10Agonum placidum 0 0 0 0.21 � 0.14 0.64 � 0.30 0.14 � 0.08 0 0.20 � 0.20 0Bembidion quadrimaculatum

oppositum0.10 � 0.05 0.10 � 0.05 0.10 � 0.05 0.26 � 0.15 1.45 � 0.57 0.71 � 0.14 0.20 � 0.12 0.20 � 0.20 0.10 � 0.10

Bembidion rapidum 0 0 0 0.14 � 0.08 1.52 � 1.04 0.29 � 0.12 0.20 � 0.12 0.10 � 0.10 0.10 � 0.10Elaphropus spp. 0 0 0 6.52 � 0.54 6.16 � 2.17 7.95 � 2.08 0 0 0Microlestes linearis (LeConte) 0 0 0 0.90 � 0.39a 1.57 � 0.29b 0.50 � 0.27a 0 0.20 � 0.12 0.10 � 0.10Stenolophus sp. 0 0 0 0a 0.21 � 0.07b 0a 0 0 0Coleoptera: CoccinellidaeCoccinellid larvae 0 0.20 � 0.05 0.15 � 0.15 0.29 � 0.16 1.00 � 0.25 1.00 � 0.43 1.10 � 0.30 1.40 � 0.81 0.40 � 0.16Coccinella septempunctata 0 0 0 0 0.21 � 0.14 0.14 � 0.08 0 0.10 � 0.10 0.30 � 0.10Harmonia axyridis 0 0 0 0.14 � 0.08a 0a 0.43 � 0.08b 0.10 � 0.10 0.10 � 0.10 0.10 � 0.10Hippodamia convergens 0 0 0 0.29 � 0.12 0.57 � 0.20 0.50 � 0.24 0 0 0Scymnus rubricaudus 0a 1.38 � 0.56b 0.05 � 0.05a 0.14 � 0.08a 12.69 � 2.91b 0.07 � 0.07a 0 0 0Coleoptera: StaphylinidaeStaphylinid spp. 0.10 � 0.05 0.20 � 0.13 0.35 � 0.13 0.69 � 0.17a 5.93 � 0.13b 0.79 � 0.24a 0.60 � 0.20a 1.00 � 0.35a 2.40 � 0.28bHemiptera: Geocoridae, Nabidae, and AnthocoridaeGeocoris spp. 0 0.10 � 0.10 0.20 � 0.20 4.29 � 3.15a 33.21 � 2.28b 0.62 � 0.13a 0.90 � 0.10 1.10 � 0.38 0.30 � 0.19Nabis americoferus 0.25 � 0.13 0.15 � 0.09 0.05 � 0.05 2.36 � 1.12a 9.40 � 1.11b 1.57 � 0.55a 1.00 � 0.42 0.90 � 0.10 1.20 � 0.23Orius insidiosus 0.05 � 0.05 0.15 � 0 0.25 � 0.13 0.36 � 0.14a 4.02 � 0.40b 1.93 � 0.77a 0.70 � 0.30 0.50 � 0.25 0.70 � 0.10Hymenoptera: FormicidaeFormicid spp. 10.17 � 3.60 2.52 � 0.51 4.49 � 0.35 2.36 � 0.79a 7.52 � 0.49b 2.76 � 1.10a 2.40 � 1.28 3.10 � 1.31 2.70 � 1.39Neuroptera: ChrysopidaeChrysoperla larvae 0.05 � 0.05a 0.69 � 0.18b 0.20 � 0.13a 0.21 � 0.07a 0.14 � 0.08a 0.57 � 0.12b 0.50 � 0.30 0.20 � 0.12 0.10 � 0.10Opiliones: PhalangiidaePhalangium opilio 1.19 � 0.34 0.84 � 0.30 1.83 � 0.27 1.98 � 0.29 5.29 � 0.71 5.00 � 1.46 3.50 � 0.74 2.50 � 1.05 3.60 � 1.61Orthoptera: GryllidaeAllonemobius spp. 0.15 � 0.09 0 0.44 � 0.30 0.50 � 0.34 1.36 � 0.37 0.52 � 0.10 0.80 � 0.28 0.20 � 0.12 0.70 � 0.10Gryllus pennsylvanicus 0.69 � 0.34 0.05 � 0.05 0.25 � 0.43 0.31 � 0.13 0.48 � 0.18 0.21 � 0.21 0.40 � 0.16 0.50 � 0.10 0.60 � 0.26Total predators 24.10 � 6.21 26.17 � 5.62 22.77 � 4.09 26.05 � 4.55a 113.14 � 9.88b 49.02 � 22.25a 32.10 � 5.79 30.80 � 1.67 27.20 � 4.79

J.G.Lundgren

etal./

CropProtection

43(2013)

104e118

111

gregarius (Say) (n ¼ 5), Carabus serratus Say (n ¼ 1), Chlaenius pla-tyderus Chaudoir (n ¼ 1), C. sericeus (Forster) (n ¼ 2), C. tomentosus(Say) (n¼ 1), Cicindela punctulata Olivier (n¼ 8), Clivina bipustulata(Fabricius) (n¼ 1), Clivina impressefrons LeConte (n¼ 13), Cymindusneglectus Haldeman (n ¼ 1), C. pilosus Say (n ¼ 2), Discoderus paral-lelus (Haldeman) (n ¼ 1), Galerita janus (Fabricius) (n ¼ 1), Harpaluscaliginosus (Fabricius) (n ¼ 18), H. compar LeConte (n¼ 1), H. faunusSay (n ¼ 2), H. herbivagus Say (n ¼ 9), H. reversus Casey (n ¼ 1),Loricera pilicornis (Fabricius) (n ¼ 3), Microlestes linearis (n ¼ 11),Notiophilus sp. (n ¼ 3), Platynus decentis (Say) (n ¼ 1), Pterostichusfemoralis (Kirby) (n ¼ 14), P. melanarius (Illiger) (n ¼ 1), Scaritessubterraneus Fabricius (n ¼ 11), and Stenolophus comma (Fabricius)(n ¼ 14) (Coleoptera: Carabidae); Coleomegilla maculata (n ¼ 4),Hippodamia convergens (n ¼ 2), Hippodamia parenthesis (n ¼ 4),H. tredecimpunctata (n ¼ 1), and Harmonia axyridis (n ¼ 1) (Cole-optera: Coccinellidae); Lampyridae larvae (n ¼ 2) (Coleoptera);Geocoris sp. (n ¼ 2) (Hemiptera: Geocoridae) and N. americoferus(n ¼ 11) (Hemiptera: Nabidae); and Chrysoperla larvae (n ¼ 7)(Neuroptera: Chrysopidae). In the quadrat samples, several infre-quently collected taxa were : An unidentified Bembidiine (n ¼ 1),Calleida decora (Fabricius) (n ¼ 1), C. impressefrons (n ¼ 8); Colliurispennsylvanica (L.) (n ¼ 1), Cyclotrachelus alternans (n ¼ 1), Harpaluspensylvanicus (n¼1),Harpalus sp. (n¼1), Lebia soleaHentz (n¼1), L.pilicornis (n¼ 1),Notiophilus sp. (n¼ 1), Poecilus lucublandus (n¼ 1),Pterostichus permundus (n¼ 1), (Coleoptera: Carabidae); C. maculata(n ¼ 4), H. parenthesis (n ¼ 6), and H. tredecimpunctata (n ¼ 6)(Coleoptera: Coccinellidae); Lampyridae larvae (n¼ 10) and an adult(n ¼ 1) (Coleoptera); Reduviidae sp. (n ¼ 1) (Hemiptera); andHemerobiid sp. adult (n ¼ 3) (Neuroptera: Hemerobiidae).

Although there were differences in individual predator taxaamong the systems (Tables 1e3), there were few consistent trendsamong years in the responses of specific species to treatments. O.insidiosus and Scymnus rubricauduswere consistently highest in thecover cropped treatment across all three years of study (this latterspecies was only significantly so in the first two years of study)(Table 1). Chrysoperla adults and larvae and coccinellid larvae werehigher in the cover cropped treatment in years when oat wasplanted, but not when rye was planted (Table 1).

4.3. Weed populations

More than 20 weed taxonomic groups were present over thestudy. Two amaranths, redroot pigweed and common waterhemp(Amaranthus retroflexus and A. rudis, respectively) and commonlambsquarters (Chenopodium album) were the most abundantspecies observed throughout the study (although these specieswerenotably fewer in2006 than inother years) (Table4). Similar numbersof weed species were found among treatments, although signifi-cantly fewerweed specieswere recovered in 2006 (P¼ 0.01) than inthe other sample years (treatment: F2, 24 ¼ 0.96, P ¼ 0.40; year F2,24 ¼ 6.03, P ¼ 0.01, interaction: F4, 24 ¼ 0.73, P ¼ 0.58). Mean (SEM)species numbers per plot (pooled across years) were 5.36 � 0.80,6.27 � 0.73, 6.18 � 0.85 in the chemically intensive, cover crop, andreduced chemical management systems, respectively.

The numbers of weeds counted along the transects differedamong treatments in only one of the three study years (Table 4). Theseasonal patterns of weed community phenologies differed amongtreatments (2005: treatment: F2, 6¼ 2.11, P¼ 0.20; date: F3, 18¼ 7.00,P ¼ 0.003; interaction: F6, 18 ¼ 4.94, P ¼ 0.004; 2006: treatment: F2,8 ¼ 4.67, P ¼ 0.045; date: F2, 16 ¼ 12.75, P < 0.001; interaction: F4,16 ¼ 7.91, P ¼ 0.001; 2007: treatment: F2, 9 ¼ 2.05, P ¼ 0.18; date: F2,18 ¼ 2.03, P ¼ 0.16; interaction: F4, 18 ¼ 8.67, P < 0.001). In 2005, theseasonalweed countwas similar among treatments, although countvaried among the sample periods; also, there was a significantinteraction between treatment and sample period (Table 4). This

variability among treatments occurred during the period prior tocanopy closure, but after herbicide application to the chemicallyintensive treatment. Here, weed populations in the reduced-chemical treatment were more abundant than the other treat-ments (Fig. 3). In 2006, more weeds were counted along the tran-sects in the cover crop treatment than theother two treatmentswithmore weeds present after soybean canopy closure in the covercropped system (Fig. 3, Table 4). In 2007, therewere no season-longdifferences in weed abundance among management systems, norwas there an effect of sample period onweed abundance. However,after soybean canopyclosure, thereweremoreweeds counted in thereduced chemical treatment relative to the chemically intensivetreatment (Fig. 3, Table 4).

4.4. Yields, grain quality, and system profitability

Yield was greatest in 2005, averaging about 9000 kg per ha, andleast in 2007, averaging about 2500 kg per ha. A 6-wk droughtbeginning on June 15 reduced yields in 2007 relative to the first twostudyyears. Yielddifferences among themanagement systemswereonly observed in 2006, where yields in the treatments were rankedas reduced chemical¼ chemical intensive> cover crop; in 2006, therye cover crop reduced yield by more than 50%. Management alsoaffected soybean quality. Soybean from the cover crop treatmenthad lower oil content but greater protein content than at least one ofthe two chemical treatments in 2006 and 2007. In 2005, grain in thecover crop management system had reduced protein contentcompared with the other two management regimes (Table 5).

Ordinary Least Squares regression procedure was used toinvestigate the relationship between yield and treatment. Produc-tion cost and yield per ha revenue estimates across treatments arepresented in Table 5. The global F-test for overall model significancehad a P < 0.001 (Table 6). The regression model has an adjustedR2 ¼ 0.64, indicating that 64% of the variability in yield wasexplained by the statistical model. All of the regression coefficients(bs) were statistically significant at the 5% level or lower. Regressionresults indicate that there is a 0.24 t per ha difference betweenchemically intensive and reduced chemical treatments, anda 0.68 t per ha difference between chemically intensive and covercrop treatments. The beta coefficient for the “RYE” variable is�1.25.Again, this variable captures the unique effect that rye had when itwas used as a cover crop. Using oat as a cover crop increased yieldsby 0.96 t per ha relative to the rye cover crop. Furthermore, whilethe use of a cover crop reduced yield in this study, it also reducedinput costs.

Production cost data indicate that the cover crop treatmentcomponent with conventional seed had the lowest average cost ofproduction per ha. Yield data indicate the cover crop treatmentcomponent also produced the lowest yields relative to the chemi-cally intensive and reduced chemical alternatives. In the studiedsystem, the lower cost structure associated with the cover croptreatment was not large enough to overcome the market value ofreduced yield, although the profitability of this system increased asthe study went on alongside our experience with this productionsystem. This notwithstanding, the cover crop treatment under-performed the alternatives with respect to profit (Table 5).

5. Discussion

The soybean production systems in this study had dramaticeffects on both insect and weed communities. Although efforts toreduce pesticide inputs increased beneficial species and reducedthe costs of production, the profitability of the treatments receivingpesticides (chemical intensive and reduced chemical) exceeded thecover crop treatment. In the first two years of study (Table 5), the

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118112

Table 4Weed communities found in soybeans produced under three management regimens. Data represent the mean (SEM) sum of weeds found along two diagonal cross-plot transects per plot per sample date (n ¼ 3 plots pertreatment in 2005 and n ¼ 4 in 2006 & 2007). Significant treatment effects (a ¼ 0.05) are indicated in bold italics. Within years and parameters, treatment means without letters indicate that the F-test was not significant and nopairwise testing was done. Where letters are present after means, the treatment F-test was significant (a ¼ 0.05), and means followed by the same letter are not significantly different (LSD multiple comparison test). Additionaltaxa that were infrequently collected are listed in the results section.

Species Common name 2005 2006 2007

Chemicalintensive

Cover crop Reducedchemical

Chemicalintensive

Cover crop Reducedchemical

Chemicalintensive

Cover crop Reducedchemical

Amaranthus sp. Redroot pigweed &commonwaterhemp

1.13 � 1.03 0.65 � 0.53 2.23 � 1.12 0.06 � 0.06 0.08 � 0.05 0.08 � 0.08 4.94 � 1.12 7.11 � 2.23 15.03 � 5.75

Chenopodium album L. Common lambsquarters 0.13 � 0.04a 0.19 � 0.04a 1.65 � 0.54b 0.36 � 0.21 0.92 � 0.70 0.47 � 0.25 2.25 � 0.67 1.89 � 0.63 4.97 � 2.57Helianthus sp. Wild sunflower 0.29 � 0.06 0.25 � 0.11 0.36 � 0.27 0.47 � 0.27 2.25 � 1.20 0.39 � 0.22 0 0.11 � 0.08 0.11 � 0.05Malva neglecta Wallr. Common mallow 0.13 � 0.10 0.40 � 0.33 0 0 0 0 0.08 � 0.05 0.03 � 0.03 0.28 � 0.13Acer sp. Maple seedling 0 0.11 � 0.11 0.40 � 0.31 0.03 � 0.03 0 0.03 � 0.03 0 0 0Trifolium repens L. White clover 0.15 � 0.09 0.08 � 0.06 0.06 � 0.06 0 0 0 0.03 � 0.03 0.03 � 0.083 0Zea mays L. Volunteer corn 0 0 0 0 0.14 � 0.08 0.06 � 0.06 0.11 � 0.05 0.11 � 0.08 0.17 � 0.11Poaceae spp. Unidentified grasses 0.11 � 0.08 0 0.17 � 0.09 0.03 � 0.03 0.11 � 0.05 0.08 � 0.05 0.11 � 0.06a 0.81 � 0.18b 0.08 � 0.08aPolygonum convolvulus L. Wild buckwheat 0.02 � 0.02 0.02 � 0.02 0.15 � 0.15 0.03 � 0.03 0.06 � 0.06 0 0.06 � 0.06 0 0.06 � 0.06Cirsium spp. Unidentified thistle

species0.06 � 0.04 0.02 � 0.02 0.08 � 0.02 0 0 0 0.03 � 0.03 0 0

Solanum ptychanthumDunal.ex DC.

Eastern blacknightshade 0 0.02 � 0.02 0.13 � 0.07 0.08 � 0.08 0.11 � 0.05 0.33 � 0.10 0 0.06 � 0.03 0.06 � 0.06

Oxalis dillenii Jacq. Gray-green woodsorrel 0 0.06 � 0.04 0.06 � 0.06 0 0.03 � 0.03 0 0.25 � 0.08b 0a 0.14 � 0.05abThlaspi arvense L. Field pennycress 0.04 � 0.02 0 0.08 � 0.08 0 0 0 0 0 0.06 � 0.03Datura stramonium L. Jimsonweed 0.02 � 0.02 0 0.02 � 0.02 0 0 0 0 0 0Portulaca oleracea L. Common purselane 0 0.02 � 0.02 0.02 � 0.02 0 0 0.03 � 0.03 0.03 � 0.03 0.14 � 0.14 0.17 � 0.03Solanum rostratum Dunal Buffalobur 0 0 0.082 � 0.02 0 0 0 0 0 0Polygonum arenastrum

Jord. Ex BoreauCommon knotweed 0 0.02 � 0.02 0 0 0 0 0 0 0

Ambrosia artemisiifolia L. Common ragweed 0 0 0 0 0.03 � 0.03 0 0 0 0Taraxacum officinale

G.H. Weber ex WiggersDandelion 0 0 0 0 0.03 � 0.03 0.03 � 0.03 0.14 � 0.11 0.08 � 0.05 0.25 � 0.07

Brassica kaber (DC.)L.C. Wheeler

Wild mustard 0 0 0 0.08 � 0.08 0.03 � 0.03 0.06 � 0.06 0 0 0

Unidentified seedling 0.27 � 0.18 0.31 � 0.11 0.52 � 0.25 0.06 � 0.03 0.11 � 0.05 0.06 � 0.03 0.17 � 0.03a 0.39 � 0.10b 0.08 � 0.03aUnknown sp. 0.02 � 0.02 0.02 � 0.02 0 0.03 � 0.03 0 0.06 � 0.06 0 0 0Total weeds 2.36 � 1.33 2.17 � 1.13 5.94 � 1.84 1.22 � 0.53a 3.89 � 0.83b 1.67 � 0.05a 8.19 � 1.44 10.78 � 2.87 21.47 � 7.87

J.G.Lundgren

etal./

CropProtection

43(2013)

104e118

113

chemically intensive and reduced chemical treatments had verysimilar profits, indicating that reducing herbicide inputs to a singleapplication before canopy closure was sufficient for managingtypical weed pressure. In 2007, one plot in each of the reducedchemical and cover cropped treatments had extraordinary weedseed input (the farmer had fed cows in this area of the field) thatcould only be overcome with more intensive herbicide use; theother reduced chemical plots had equivalent yields to the chemicalintensive treatment in 2007 (see footnote in Table 5). There wereseveral deficiencies (discussed below) in our management of thecover crop treatment that made this the least profitable of the threesystems. If additional revenue could be added to this cover cropproduction system, either through gaining organic certification ofthe grain or through reducing the impact of the cover crop onyields, growers could take advantage of the lower costs ofproduction and favorable habitats provided by this treatment tocreate more sustainable, long-term solutions for pest management.

Soybean aphid populations differed among the treatments, withinsecticide applications having the greatest influence on relativeaphid populations (Fig. 1). In all three seasons, a single applicationof lambda cyhalothrin (but not esfenvalerate) was sufficient toreduce mounting aphid infestations below 250 per plant for theremainder of the season. When the oat cover crop was removedfrom the field early in the season (2005 & 2007), aphid populationsexceeded the EIL in all plots by season’s end. When the rye covercrop was allowed to persist throughout the season, aphid pop-ulations were kept below 50 aphids per plant throughout theseason. This same pattern was observed with the ground cover inSchmidt et al. (2007); in this case, an alfalfa living mulch groundcover was allowed to persist in the understory of the soybean crop,and aphid populations were diminished as a result. But also likeSchmidt et al. (2007), we found that when the rye was allowed topersist season-long, yields of soybean were significantly reducedby approximately 67% (Table 5), making this strategy unappealingfor farmers. In a recent survey of organic soybeans preceded bywinter rye cover crops, Koch et al. (2012) found that aphids werereduced through cover-cropping and yields were unaffected by thecover crop, but in this study yields were exceptionally low even inthe untreated control (<1.35 t per ha). Finally, incorporating wheatinto soybeans as an intercrop successfully reduced potato leaf-hopper, while having no effects on soybean yield (Hammond andJeffers, 1990; Miklasiewicz and Hammond, 2001). All of this takentogether suggests that cover crops and ground covers havepotential for reducing soybean aphids, but care must be taken toensure that the cover crop is managed in ways that maximizesoybean yields.

Bean leaf beetle populations differed among managementsystems, but only during the first year of study when the pest wasparticularly abundant (Table 2). The first study year experiencedbeetle populations that were 10-fold higher in the reduced chem-ical and cover crop systems than in 2006 and 2007 (representinga 7-year high for this research site). Numbers remained between 10and 100 beetles (per 100 sweeps) from July 22 until August 26,2005, which corresponds to the adult stage of the 1st generation(i.e. the progeny of the overwintering generation) of this pest ineastern South Dakota (Hammack et al., 2010; Riedell et al., 2011).Yield correlations for this pest based on sweeps suggest that thesepopulations did not approach economically damaging levels in thereduced chemical and cover crop treatments in 2005 (Hammacket al., 2010). Aphid-targeted insecticide inputs in the chemicaltreatment had the added benefit of reducing bean leaf beetlepopulations in 2005; this benefit of insecticides targeted at other,more serious insect pests of soybeans is found in other regions aswell (Johnson et al., 2008; Musser et al., 2012). In summary, beanleaf beetle populations are affected by insecticides that target

Fig. 3. Weed seasonal dynamics in soybeans produced under three managementregimens. Weed populations were measured every 0.5 m along two 32 m transects ineach plot (n ¼ 3 in 2005, 4 in 2006 & 2007). Populations were recorded prior toherbicide applications, after herbicide applications but prior to soybean canopyclosure, and after canopy closure. In 2005, an additional observation was recorded afterherbicides were applied to the chemically intensive treatment, but before herbicideswere sprayed in the reduced chemical treatment. Bars represent mean (SEM) weednumber per plot, and bars capped with different letters are significantly different(a ¼ 0.05).

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118114

soybean aphids, but bean leaf beetle populations were noteconomically threatening evenwhen insecticides were relaxed andvegetation was diversified by weeds or cover crops.

A diverse natural enemy communitywas discovered in soybeansof the Northern Great Plains, but only the foliar-dwelling predatorcommunity differed among the management systems (Tables 1e3,Fig. 2). Predators are commonly shown to be an important factorsuppressing soybean aphid population growth (Fox et al., 2004;Rutledge et al., 2004; Desneux et al., 2006; Brosius et al., 2007;Rhainds et al., 2007; Schmidt et al., 2007; Costamagna et al., 2008).Over the 3 yr study, we collected at least 83 species of predators;this number would have increased if we had identified the For-micidae, Staphylinidae and Araneae to species level. Foliar predatorcommunities vary depending on the region of study, and theirrelative abundance and community composition may havechanged since the invasion of the soybean aphid (Schmidt et al.,2008). In Canada, Michigan, and Iowa, the main predators foundin association with soybean aphids are coccinellids (Rhainds et al.,2007; Costamagna et al., 2008; Schmidt et al., 2008), often domi-nated by H. axyridis later in the growing season (Rutledge et al.,2004). But natural enemy communities in the Northern GreatPlains (NE and SD) are dominated by O. insidiosus (Anthocoridae),N. americoferus (Nabidae), and spiders (Brosius et al., 2007;Seagraves and Lundgren, 2012). The current work substantiatesthat while coccinellids were a consistent presence in our fields(especiallyH. axyridis late in the season), predatory hemipterans (O.insidiosus and N. americoferus) dominated the predator guild livingin the soybean foliage (but note that we did not describe the spidercommunity in this study). At least O. insidiosus is regarded as one ofthe primary predators of the soybean aphid, especially in delayingearly season build-up of aphid populations and reducing smalleraphid populations (Rutledge and O’Neil, 2005; Desneux et al., 2006;Costamagna et al., 2008). With these observations, it is difficult tosay which of the predators are having the greatest impact on aphidpopulations, but in numerical terms our work echoes previousstudies that O. insidiosus and H. axyridis are likely important, andunderscores that N. americoferus is on the list of potentiallyimportant biological control agents of soybean aphid. The soilsurface in soybeans has a diverse community of predatorsthroughout the soybean producing region, often reportedly domi-nated by spiders and carabid beetles (Rutledge et al., 2004;Gardiner et al., 2010). Our work suggests that this may be an artifactof the choice of sampling procedure. We do not deny that carabidsand spiders are an important and diverse component of the naturalenemy guild, but our data suggest that their predominance is likelyoverinflated due to pitfall sampling.We used pitfall traps to capturedozens of carabid species in soybeans (Table 2 and Results Section),

Table 6Regression analysis of effect of three pest management systems (chemically inten-sive, reduced chemical, and cover crop) on soybean yields over a three year, repli-cated experiment.

Dependent variable: soybean yield (t per ha)

Mean 2.86 Obs. 92SSR 56,53 Mean SSR 11.30SSE 28.57 Mean SSE 0.33F 34.03 P < 0.001R2 0.66 Adj R2 ¼ 0.64

Independent variables: Coefficient Std. Error T-stat.

Constant 2.71 0.08 33.77Reduced chemical �0.24 0.12 �1.97Cover crop �0.68 0.20 �3.44Year 3 0.91 0.14 6.35Year 1 0.82 0.11 7.24RYE �1.25 0.24 �5.23

Table

5Yield

andgrainqu

alityforsoyb

eansproducedunder

threeman

agem

entreg

imen

s.Datapresentedismea

n(SEM

)per

plot(n¼3plots

per

trea

tmen

tin20

05an

dn¼4in

2006

&20

07).Sign

ificanttreatmen

teffects

(a¼0.05

)are

indicated

inbo

lditalics.W

ithin

yearsan

dparam

eters,trea

tmen

tmea

nswithou

tlettersindicatethat

theF-test

was

not

sign

ificantan

dnopairw

isetestingwas

don

e.W

herelettersarepresentaftermea

ns,thetrea

tmen

tF-test

was

sign

ificant( a

¼0.05

),an

dmea

nsfollo

wed

bythesameletter

arenot

sign

ificantlydifferent(LSD

multiple

comparison

test).

2005

2006

2007

Chem

ical

intensive

Cov

ercrop

Red

ucedch

emical

Chem

ical

intensive

Cov

ercrop

Red

ucedch

emical

Chem

ical

intensive

Cov

ercrop

Red

ucedch

emical

Dry

matter(%)

91.76�

0.22

91.63�

0.36

91.02�

1.08

94.17�

0.48

95.04�

0.29

93.90�

0.26

92.24�

0.61

91.72�

0.95

91.70�

0.91

Oil(%)

19.26�

0.06

19.89�

0.01

19.57�

0.24

19.23�

0.24

a18

.54�

0.15

a20

.03�

0.27

b19

.52�

0.03

a18

.84�

0.18

a19

.36�

0.03

bProtein(%)

38.68�

0.19

c37

.46�

0.16

a38

.26�

0.31

b37

.19�

0.68

a39

.19�

0.42

b36

.09�

0.71

a38

.63�

0.15

a39

.54�

0.13

b38

.95�

0.25

abYield

(tha�

1)

10.47�

0.59

8.69

�0.98

8.56

�0.44

6.89

�0.16

2.45

�0.57

7.40

�0.24

3.55

�0.12

2.29

�0.77

2.47

�0.89

Costs

(ha�

1)

$296

.82

$213

.00

$191

.50

$269

.83

$104

.38

$182

.83

$188

.04

$161

.78

$179

.64

Profi

t(ha�

1)

$422

.52�

40.10

$382

.73�

67.58

$405

.84�

37.13

$285

.47�

15.72

$89.89

�45

.17

$402

.72�

20.11

$115

6.97

�57

.01

$724

.08�

300.03

a$7

89.39�

355.01

a

aEa

chof

thesetrea

tmen

tslost

aplotto

unch

aracteristically

highwee

dpressure.W

hen

this

replic

ateis

omittedfrom

thean

alysis,themea

nprofits

were$1

001an

d$1

129per

hain

theco

vercrop

andreducedch

emical

trea

tmen

ts.

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118 115

although spiders were not as abundant in our systems as in surveysfrom other regions (Gardiner et al., 2010). However, when weexamined actual densities of predators, carabid beetles were muchless abundant than spiders, ants, and harvestmen in our systems(Table 3). These observations on actual densities of predators aresupportive of other reports involving quadrat samples (Lundgrenet al., 2006; Lundgren and Fergen, 2010). Regardless, employingmultiple sample methods helps give a better understanding ofpredator diversity within soybeans of the Northern Great Plains.

The foliar-dwelling predator community was the only oneaffected by the management treatments. This observation mayhave resulted because there was more prey present in the covercropped treatment and because insecticides likely reduced pred-ator populations in the chemical treatments. Other work has shownthat insecticide use in soybeans (Ohnesorg et al., 2009; Seagravesand Lundgren, 2012) and other crops (Marvier et al., 2007;Wolfenbarger et al., 2008) can reduce some or all natural enemyspecies present within a habitat. O. insidiosus and S. rubricauduswere consistently more abundant in the cover crop system than inthe other two management systems (Table 1). O. insidiosus is animportant predator of soybean thrips and the soybean aphid(Brosius et al., 2007; Butler and O’Neil, 2008; Desneux and O’Neil,2008; Harwood et al., 2009). This omnivorous bug also has strongpreferences for specific plant species as oviposition sites, andincreased vegetation diversity has been shown to increaseO. insidiosus populations (Lundgren et al., 2008, 2009b; Seagravesand Lundgren, 2010). Insecticides aimed at aphid managementalso adversely affect O. insidiosus (Seagraves and Lundgren, 2012).We believe that the provision of preferred non-crop vegetation andpossibly food diversity, as well as reduced insecticide use, in largepart explain the increased numbers of O. insidiosus in the cover cropsystem. Very little is known about the natural history ofS. rubricaudus, but the total predator population peaks observed inthe cover crop system in late June to early July of 2005 and 2006were the result of mass emergence of S. rubricaudus adults (Fig. 2).The congeners S. louisianae J. Chapin and S. babai Sasaji have beenidentified as predators of soybean aphids in Kentucky and Asia,respectively (Brown et al., 2003; Liu et al., 2004). We have consis-tently captured this predator in field plots of eastern South Dakotathat were preceded by a spring grass cover crop (Lundgren andFergen, 2010), and identify this species as meriting further atten-tion from the perspective of pest management.

Weed phenology was affected by management, but only atspecific times during the season. At least 20 weed species werecharacterized during the 3 yr study; however, this number is anunderestimate, as grasses, small seedlings with less than three trueleaves, and thistles were not identified to species level. The mostabundant weeds in these systems were common lambsquarters,redroot pigweed, and common waterhemp (Table 4). The chemi-cally intensive treatment had some of the lowest weed abundancesobserved (Table 4, Fig. 3), although the only significant differencesamong treatments in seasonal weed abundances occurred in 2006(the cover crop system had significantly higher weed densities thanthe other two systems). Cover crops (and the management asso-ciated with removing them) suppressed weed populations in twoof the three growing seasons to similar levels to those observed inthe systems managed exclusively with herbicides. In 2006, whenrye was left in the plots throughout the season, greater weednumbers were present in the cover crop system than the othersystems (especially after the crop canopy closed). This wassurprising, as rye produces allelochemicals known to suppresscompetition from other plants (Mwaja et al., 1995; Reberg-Hortonet al., 2005). But this was also the only year of study when thecover cropwas not killed with herbicides, andwe believe that these

herbicide applications suppressed weeds that were adapted togrow in concert with the cover crops. Weed populations weremoreabundant in the reduced chemical system over the chemicallyintensive system in 2005 and 2007, but only in one of the timeperiods sampled. Specifically, weeds in the reduced chemicalsystemwere particularly high just prior to the crop canopy closureand before herbicides were applied in 2005, and after the canopyhad closed in 2007 (Fig. 3). The end result is that the reducedchemical treatment produced greater weed abundances duringportions of the growing season, but under normal circumstancesthe timing and abundance of these weed populations in thistreatment had minimal effects on soybean profitability whencompared to the weed-free, chemically intensive treatment.

There are several ways that producers could increase the prof-itability of the cover crop system to make this competitive or evensuperior to the chemically intensive treatment. First, our annualobservations always represented the first year of cover crop use. Itis common to see a yield drag associated with first-year cover cropuse, but in subsequent years yields typically improve (Osborneet al., 2008). Also, we found that managing the cover crop moreefficiently (e.g., killing the cover crop entirely prior to soybeanestablishment) was key to improving yields of the subsequentsoybean crop. Although we found mowing to be ineffective andultimately used herbicides to kill the cover crop in our study,innovative non-chemical technologies such as roller/crimpers offerorganically compatible solutions to eliminating cover crops fromfarmland (Mirsky et al., 2009; Davis, 2010). Although cover croppedconventional soybean yields could feasibly be as profitable assoybeans managed chemically, producers that target the organicmarket could have even greater profits than chemically managedsoybeans due to increased organic premiums. For example, if thecover crop approach was able to earn an organic premium(approximately $335.76 per t; report NW_GR_113, http://search.ams.usda.gov/mnsearch/mnsearch.aspx), then even the lowestsoybean yields from the 2006 cover crop treatment would havebeenmore profitable than the chemical intensive treatment. Sufficeit to say that we feel that the cover crop treatment has substantialpotential as a long-term, sustainable form of pest management ifproduction can be amended so that yields are improved, largelybecause of the conservation benefits that this approach has fornatural enemies. Natural enemies of soybean aphids save producersbillions of dollars every year (Landis et al., 2008), and the increasingreliance on insecticides in corn and soybean systems observed inSouth Dakota (Fausti et al., in press) and the rest of the United States(Meehan et al., 2011) challenges the ability of these beneficialspecies to contribute their services to farmers.

Acknowledgments

This work would not have been possible without the steadfastassistance of Janet Fergen and Eric Beckendorf, and Carson Dinger,Kelly Heitkamp, Mallory Johnson, Matt Jones, Lacey Kruse, MalissaMayer, and Richard White in collecting and processing the insectsand collating the data. Doris Lagos and David Voegtlin helped toidentify soybean aphids in tile traps. Emmanuel Opuku assistedwith locating historical cost information on the inputs. Max Pra-vacek assisted in managing the field plots. Beth Choate (USDA-ARS)and Kelley Tilmon (South Dakota State University) provided helpfulcomments on an earlier draft of this manuscript. Mention of tradenames or commercial products in this publication is solely forthe purpose of providing specific information and does notimply recommendation or endorsement by the U.S. Department ofAgriculture.

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118116

Appendix 1. Cost data. All data in the model was calculated asprice or cost per ha.a

References

Andow, D.A., 1991. Vegetational diversity and arthropod population response. Annu.Rev. Entomol. 36, 561e586.

Beckendorf, E.A., Catangui,M.A., Riedell,W.E., 2008. Soybean aphid feeding injuryandsoybean yield, yield components, and seed composition. Agron. J. 100, 237e246.

Bradshaw, J.D., Rice, M.E., Hill, J.H., 2008. Evaluation of management strategies forbean leaf beetles (Coleoptera: Chrysomelidae) and bean pod mottle virus(Comoviridae) in soybean. J. Econ. Entomol. 101, 1211e1227.

Broad, S.T., Lisson, S.N., Mendham, N.J., 2009. Agronomic and gross margin analysisof an insect pest suppressive broccoli cropping system. Agric. Syst. 102, 41e47.

Brosius, T.R., Higley, L.G., Hunt, T.E., 2007. Population dynamics of soybean aphidand biotic mortality at the edge of its range. J. Econ. Entomol. 100, 1268e1275.

Brown, C.G., Sharkey, M.J., Johnson, D.W., 2003. Bionomics of Scymnus (Pullus)louisianae J. Chapin (Coleoptera: Coccinellidae) as a predator of the soybeanaphid, Aphis glycines Matsumura (Homoptera: Aphididae). J. Econ. Entomol. 96,21e24.

Butler, C.D., O’Neil, R.J., 2008. Voracity and prey preference of insidious flower bug(Hemiptera: Anthocoridae) for immature stages of soybean aphid (Hemiptera:Aphididae) and soybean thrips (Thysanoptera: Thripidae). Environ. Entomol. 37,964e972.

Byamukama, E., Robertson, A.E., Nutter, F.W.J., 2011. Quantifying the within-fieldtemporal and spatial dynamics of Bean pod mottle virus in soybean. Plant Dis.95, 126e136.

Clark, A.J., Perry, K.L., 2002. Transmissibility of field isolates of soybean viruses byAphis glycines. Plant Dis. 86, 1219e1222.

Costamagna, A.C., Landis, D.A., Brewer, M.J., 2008. The role of natural enemy guildsin Aphis glycines suppression. Biol. Control 45, 368e379.

Davis, A.S., 2010. Cover-crop roller-crimper contributes to weed management in no-till soybean. Weed Sci. 58, 300e309.

Desneux, N., O’Neil, R.J., 2008. Potential of an alternative prey to disrupt predationof the generalist predator, Orius insidiosus, on the pest aphid, Aphis glycines, viashort-term indirect interactions. Bull. Entomol. Res. 98, 631e639.

Desneux, N., O’Neil, R.J., Yoo, H.J.S., 2006. Suppression of population growth of thesoybean aphid, Aphis glycines Matsumura, by predators: the identification ofa key predator, and the effects of prey dispersion, predator density andtemperature. Environ. Entomol. 35, 1342e1349.

Fausti, S.W., McDonald, T.M., Lundgren, J.G., Li, J., Keating, A.R., Catangui, M.A.Insecticide use and crop selection in regions with high GM adoption rates.Renew. Agr. Food Syst., in press.

Fox, T.B., Landis, D.A., Cardoso, F.F., DiFonzo, C.D., 2004. Predators suppress Aphisglycines Matsumura population growth in soybean. Environ. Entomol. 33,608e618.

Gardiner, M.M., Landis, D.A., Gratton, C., Schmidt, N., O’Neal, M., Mueller, E.,Chacon, J., Heimpel, G.E., 2010. Landscape composition influences the activitydensity of Carabidae and Arachnida in soybean fields. Biol. Control 55, 11e19.

Griffen, M.L., Yeargan, K.V., 2002. Oviposition site selection by the spotted lady-beetle Coleomegilla maculata (Coleoptera: Coccinellidae): choices among plantspecies. Environ. Entomol. 31, 107e111.

Hammack, L., Pikul Jr., J.L., West, M.S., 2010. Phenology and abundance of beanleaf beetle (Coleoptera: Chrysomelidae) in Eastern South Dakota on alfalfaand soybean relative to tillage, fertilization, and yield. Environ. Entomol. 39,727e737.

Hammond, R.B., Jeffers, D.L., 1990. Potato leafhopper (Homoptera: Cicadellidae)populations on soybean intercropped into winter wheat. Environ. Entomol. 19,1810e1819.

Harwood, J.D., Yoo, H.J.S., Rowley, D.L., Greenstone, M.H., O’Neil, R.J., 2009. Differ-ential impact of adults and nymphs of a generalist predator on an exoticinvasive pest demonstrated by molecular gut-content analysis. Biol. Invasions11, 895e903.

Heap, I., 2012. International Survey of Herbicide Resistant Weeds. www.weedscience.org.

Hodgson, E.W., Koch, R.L., Ragsdale, D.W., 2005. Pan trapping for soybean aphid(Homoptera: Aphididae) in Minnesota soybean fields. J. Entomol. Sci. 40,409e419.

Irwin, M.E., Goodman, R.M., 1981. Ecology and control of soybean mosaic virus. In:Maramorosch, K., Harris, K.F. (Eds.), Plant Diseases and Their Vectors: Ecologyand Epidemiology. Academic Press, New York, pp. 181e222.

Johnson, K.D., O’Neal, M.E., Bradshaw, J.D., Rice, M.E., 2008. Is preventative,concurrent management of the soybean aphid (Hemiptera: Aphididae) andbean leaf beetle (Coleoptera: Chrysomelidae) possible? J. Econ. Entomol. 101,801e809.

Koch, R.L., Porter, P.M., Harbur, M.M., Abrahamson, M.D., Wyckhuys, K.A.G.,Ragsdale, D.W., Koch, K., Sezen, Z., Heimpel, G.E., 2012. Response of soybeaninsects to an autumn-seeded rye cover crop. Environ. Entomol. 41, 750e760.

Landis, D.A., Gardiner, M.M., van der Werf, W., Swinton, S.M., 2008. Increasing cornfor biofuel production reduces biocontrol services in agricultural landscapes.Proc. Natl. Acad. Sci. U S A 105, 20552e20557.

Letourneau, D.K., Armbrecht, I., Rivera, B.S., Lerma, J.M., Carmona, E.J., Daza, M.C.,Escobar, S., Galindo, V., Gutiérrez, C., López, S.D., Mejía, J.L., Rangel, A.M.A.,Rangel, J.H., Rivera, L., Saavedra, C.A., Torres, A.M., Trujillo, A.R., 2011. Does plantdiversity benefit agroecosystems? A synthetic review. Ecol. Appl. 21, 9e21.

Liu, J., Wu, K., Hopper, K.R., Zhao, K., 2004. Population dynamics of Aphis glycines(Homoptera: Aphididae) and its natural enemies in soybean in northern China.Ann. Entomol. Soc. Am. 97, 235e239.

Lundgren, J.G., Fergen, J.K., 2010. The effects of a winter cover crop on Diabroticavirgifera (Coleoptera: Chrysomelidae) populations and beneficial arthropodcommunities in no-till maize. Environ. Entomol. 39, 1816e1828.

Lundgren, J.G., Fergen, J.K., 2011. Enhancing predation of a subterranean insect pest:a conservation benefit of winter vegetation in agroecosystems. Appl. Soil Ecol.51, 9e16.

Lundgren, J.G., Fergen, J.K., Riedell, W.E., 2008. The influence of plant anatomy onoviposition and reproductive success of the omnivorous bug, Orius insidiosus.Anim. Behav. 75, 1495e1502.

Lundgren, J.G., Gassmann, A.J., Bernal, J.S., Duan, J.J., Ruberson, J.R., 2009a. Ecologicalcompatibility of GM crops and biological control. Crop Prot. 28, 1017e1030.

Lundgren, J.G., Riedell, W.E., 2008. Soybean nitrogen relations and root character-istics after Cerotoma trifurcata (Coleoptera: Chrysomelidae) larval feedinginjury. J. Entomol. Sci. 43, 107e116.

Lundgren, J.G., Shaw, J.T., Zaborski, E.R., Eastman, C.E., 2006. The influence oforganic transition systems on beneficial ground-dwelling arthropods andpredation of insects and weed seeds. Renew. Agr. Food Syst. 21, 227e237.

Lundgren, J.G., Wyckhuys, K.A.G., Desneux, N., 2009b. Population responses by Oriusinsidiosus to vegetational diversity. BioControl 54, 135e142.

Mabry, T.R., Hobbs, H.A., Steinlage, T.A., Johnson, B.B., Pedersen, W.L., Spencer, J.L.,Levine, E., Isard, S.A., Domier, L.L., Hartman, G.L., 2003. Distribution of leaf-feeding beetles and Bean pod mottle virus (BPMV) in Illinois and trans-mission of BPMV in soybean. Plant Dis. 87, 1221e1225.

Macedo, T.B., Bastos, C.S., Higley, L.G., Ostlie, K.R., Madhavan, S., 2003. Photosyn-thetic responses of soybean to soybean aphid (Homoptera: Aphididae) injury.J. Econ. Entomol. 96, 188e193.

Marvier, M., McCreedy, C., Regetz, J., Kareiva, P., 2007. A meta-analysis of effects ofBt cotton and maize on nontarget invertebrates. Science 316, 1475e1477.

Meehan, T.D., Werling, B.P., Landis, D.A., Gratton, C., 2011. Agricultural landscapesimplification and insecticide use in the Midwestern United States. Proc. Natl.Acad. Sci. U S A 101, 11500e11505.

Miklasiewicz, T.J., Hammond, R.B., 2001. Density of potato leafhopper (Homoptera:Cicadellidae) in response to soybean-wheat cropping systems. Environ. Ento-mol. 30, 204e214.

Mirsky, S.B., Curran, W.S., Mortensen, D.A., Ryan, M.R., Shumway, D.L., 2009. Controlof cereal rye with a roller/crimper as influenced by cover crop phenology.Agron. J. 101, 1589e1596.

Input/output Price Reference

2005 2006 2007

Chemicalcosts

Metolachlor $410 L�1 $407 L�1 e 1Glyphosate $128 L�1 $111 L�1 $108 L�1 1Esfenvalerate e $384 L�1 e 1Lambda-cyhalothrin

$1425 L�1 $1459 L�1 $1406 L�1 2

Seed costs Soybeanseed(non-GM)

$495 t�1 $554 t�1 $882 t�1 3

Soybeanseed (GM)

$1516 t�1 $1378 t�1 $1332 t�1 3

Technologyfees

$551 t�1 $551 t�1 $551 t�1 3

Rye seed e $38.45 ha�1 e 4Oat seed 34.14 ha�1 e 34.14 ha�1 4

Miscellaneouscosts

Cover cropplanting

$19.77 ha�1 $22.24 ha�1 $24.71 ha�1 3

Mowing $12.36 ha�1 $13.59 ha�1 $14.83 ha�1 3Chemicalapplication

$12.36 ha�1 $13.59 ha�1 $14.83 ha�1 3

Soybean yield price $248 t�1 $277 t�1 $441 t�1 5

1Agricultural statistics 2005e2008.2AgFirst Farmers Cooperative, Brookings, SD, USA, 57006.3SDSU Extension (field personnel) indicate seed cost twice bean price.4Milborn Seeds, Brookings, SD, USA, 57006.5South Dakota Agriculture 2009, South Dakota Agricultural Statistics Service,Bulletin No. 69, June 2009.

a The text indicates a planting rate of 370,000 seeds per hectare. There are150,000 soybean seeds per bushel. So it is assumed the application rate on onebushel per acre for the non GM treatment. It is also assumed that the price of nonGM soybean seed is twice the market price for a bushel of soybeans. South DakotaState University Cooperative Extension Service provided estimates of GM seed costand technology fees. Application costs and mowing costs were determined by thenumber of times an activity was reported.

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118 117

Mortensen, D.A., Egan, J.F., Maxwell, B.D., Ryan, M.R., Smith, R.G., 2012. Navigatinga critical juncture for sustainable weed management. BioScience 62, 75e84.

Musser, F.R., Knighten, K.S., Smith, J.F., Catchot, A.L., 2012. Pyrethroid insecticidetolerance in bean leaf beetle, Certoma trifurcata, in the Mississippi Delta. PestManag. Sci. 68, 658e662.

Mwaja, V.N., Masiunas, J.B., Weston, L.A., 1995. Effects of fertility on biomass,phytotoxicity, and allelochemical content of cereal rye. J. Chem. Ecol. 21, 81e96.

Ohnesorg, W.J., Johnson, K.D., O’Neal, M.E., 2009. Impact of reduced-risk insecti-cides on soybean aphid and associated natural enemies. J. Econ. Entomol. 102,1816e1826.

Osborne, S.L., Schumacher, T.E., Humburg, D.S., 2008. Evaluation of cover crops toincrease corn emergence, yield and field trafficability. Agric. J. 3, 397e400.

Ragsdale, D.W., Landis, D.A., Brodeur, J., Heimpel, G.E., Desneux, N., 2011. Ecologyand management of the soybean aphid in North America. Annu. Rev. Entomol.56, 378e399.

Ragsdale, D.W., McCornack, B.P., Venette, R.C., Potter, B.D., MacRae, I.V.,Hodgson, E.W., O’Neal, M.E., Johnson, K.D., O’Neil, R.J., DiFonzo, C.D., Hunt, T.E.,Glogoza, P.A., Cullen, E.M., 2007. Economic threshold for soybean aphid(Hemiptera: Aphididae). J. Econ. Entomol. 100, 1258e1267.

Ragsdale, D.W., Voegtlin, D.J., O’Neil, R.J., 2004. Soybean aphid biology in NorthAmerica. Ann. Entomol. Soc. Am. 97, 204e208.

Reberg-Horton, S.C., Burton, J.D., Danehower, D.A., Ma, G.Y., Monks, D.W.,Murphy, J.P., Ranells, N.N., Williamson, J.D., Creamer, N.G., 2005. Changes overtime in the allelochemical content of ten cultivars of rye (Secale cereale L.).J. Chem. Ecol. 31, 179e193.

Rhainds, M., Roy, M., Daigle, G., Brodeur, J., 2007. Toward management guidelinesfor the soybean aphid in Quebec. I. Feeding damage in relationship to sea-sonality of infestation and incidence of native predators. Can. Entomol. 139,728e741.

Riedell, W.E., Catangui, M.A., 2006. Greenhouse studies of soybean aphid (Hemi-ptera: Aphididae) eff ects on plant growth, seed yield and composition. J. Agr.Urban Entomol. 23, 225e235.

Riedell, W.E., Catangui, M.A., Beckendorf, E.A., 2009. Nitrogen fixation, ureide, andnitrate accumulation responses to soybean aphid injury in Glycine max. J. PlantNutr. 32, 1674e1686.

Riedell, W.E., Osborne, S.L., Lundgren, J.G., Pikul Jr., J.L., 2011. Nitrogen fertilizermanagement effects on soybean nitrogren components and bean leaf beetlepopulations. Agron. J. 103, 1432e1440.

Rutledge, C.E., O’Neil, R.J., 2005. Orius insidiosus (Say) as a predator of the soybeanaphid, Aphis glycines Matsumura. Biol. Control 33, 56e64.

Rutledge, C.E., O’Neil, R.J., Fox, T.B., Landis, D.A., 2004. Soybean aphid predators andtheir use in integrated pest management. Ann. Entomol. Soc. Am. 97, 240e248.

Schmidt, N.P., O’Neal, M.E., Dixon, P.M., 2008. Aphidophagous predators in Iowasoybean: a community comparison across multiple years and samplingmethods. Ann. Entomol. Soc. Am. 101, 341e350.

Schmidt, N.P., O’Neal, M.E., Singer, J.W., 2007. Alfalfa living mulch advances bio-logical control of soybean aphid. Environ. Entomol. 36, 416e424.

Seagraves, M.P., Lundgren, J.G., 2010. Oviposition response by Orius insidiosus (Hem-iptera: Anthocoridae) to plant and prey availability. Biol. Control 55, 174e177.

Seagraves, M.P., Lundgren, J.G., 2012. Effects of neonicotinoid seed treatments onsoybean aphid and its natural enemies. J. Pest Sci. 85, 125e132.

Smith, A.N., Reberg-Horton, S.C., Place, G.T., Meijer, A.D., Arellano, C., Mueller, J.P.,2011. Rolled rye mulch for weed suppression in organic no-tillage soybeans.Weed Sci. 59, 224e231.

Tillman, G., Schomberg, H., Phatak, S., Mullinix, B., Lachnicht, S., Timper, P.,Olson, D.M., 2004. Influence of cover crops on insect pests and predators inconservation tillage cotton. J. Econ. Entomol. 97, 1217e1232.

Toepfer, S., Haye, T., Erlandson, M., Goettel, M., Lundgren, J.G., Kleespies, R.G.,Weber, D.C., Cabrera Walsh, G., Peters, A., Ehlers, R.-U., Strasser, H., Moore, D.,Keller, S., Vidal, S., Kuhlmann, U., 2009. A review of the natural enemies ofbeetles in the subtribe Diabroticina (Coleoptera: Chrysomelidae): implicationsfor sustainable pest management. Biocontrol Sci. Technol. 19, 1e65.

White, H., 1980. A heteroskedasticity-consistent covariance matrix estimator anda direct test for heteroskedasticity. Econometrica 48, 817e838.

Wolfenbarger, L.L., Naranjo, S.E., Lundgren, J.G., Bitzer, R.J., Watrud, L.S., 2008. Btcrop effects on functional guilds of non-target arthropods: a meta-analysis.PLoS One 3, e2118. http://dx.doi.org/10.1371/journal.pone.0002118.

Yoo, H.J.S., O’Neil, R.J., 2009. Temporal relationships between the generalistpredator, Orius insidiosus, and its two major prey in soybean. Biol. Control 48,168e180.

J.G. Lundgren et al. / Crop Protection 43 (2013) 104e118118